DESCRIPTION: Polymorphisms in the CAG repeat length in the AR gene have been demonstrated to correlate with risk of prostate cancer. Using blood samples collected in 1993-94 from 18,000 men in the ongoing HPFS, in combination with buccal brushings to be acquired from an additional 20,000 cohort members, there are plans to characterize risk of prostate cancer, based on CAG polymorphisms in the AR gene. Then, using a nested case-control design of approximately 1,000 projected cases from 1993-2000, the impact of specific nutritional factors on prostate cancer risk will be examined, stratifying across genetic risk. Specifically, associations of diet and nutritional status (pre-adult adiposity and attained height) will be examined with incidence of prostate cancer among men stratified by risk according to number of CAG repeats in the AR gene. Then, an assessment will be made of whether men at higher risk can potentially reduce their excess risk by decreasing their consumption of animal fat and by increasing lycopene and fiber intake, aspects of diet that have already been associated with reduced risk in the overall cohort. Non-nutritional factors that are hypothesized to influence testosterone levels will also be examined, in particular physical activity and smoking, in relation to incidence of prostate cancer among men, stratified by genetic risk according to AR characteristics. A final Aim is to acquire buccal smears for DNA analyses from HPFS participants who did not previously provide blood specimens. This will complete the establishment of a large database consisting of a) updated dietary and other exposure information over a 15 year period for 50,000 men, b) stored DNA samples (blood or buccal brushing) from approximately 38,000 of these men, c) stored plasma from 18,000 participants, and d) archived tissue from incident prostate cancers. The buccal brushings, along with providing the basis for the specific aims in this application, will allow for the rapid testing of future specific hypotheses.